代码拉取完成,页面将自动刷新
// Copyright 2016 The Cockroach Authors.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
// implied. See the License for the specific language governing
// permissions and limitations under the License.
//
// An ordered key encoding scheme for arbitrary-precision fixed-point
// numeric values based on sqlite4's key encoding:
// http://sqlite.org/src4/doc/trunk/www/key_encoding.wiki
package encoding
import (
"bytes"
"fmt"
"math"
"math/big"
"unsafe"
"github.com/pkg/errors"
"github.com/cockroachdb/apd"
)
// EncodeDecimalAscending returns the resulting byte slice with the encoded decimal
// appended to the given buffer.
//
// Values are classified as large, medium, or small according to the value of
// E. If E is 11 or more, the value is large. For E between 0 and 10, the value
// is medium. For E less than zero, the value is small.
//
// Large positive values are encoded as a single byte 0x34 followed by E as a
// varint and then M. Medium positive values are a single byte of 0x29+E
// followed by M. Small positive values are encoded as a single byte 0x28
// followed by a descending varint encoding for -E followed by M.
//
// Small negative values are encoded as a single byte 0x26 followed by -E as a
// varint and then the ones-complement of M. Medium negative values are encoded
// as a byte 0x25-E followed by the ones-complement of M. Large negative values
// consist of the single byte 0x1a followed by a descending varint encoding of
// E followed by the ones-complement of M.
func EncodeDecimalAscending(appendTo []byte, d *apd.Decimal) []byte {
return encodeDecimal(appendTo, d, false)
}
// EncodeDecimalDescending is the descending version of EncodeDecimalAscending.
func EncodeDecimalDescending(appendTo []byte, d *apd.Decimal) []byte {
return encodeDecimal(appendTo, d, true)
}
func encodeDecimal(appendTo []byte, d *apd.Decimal, invert bool) []byte {
if d.IsZero() {
// Negative and positive zero are encoded identically. Only nonsorting
// decimal encoding can retain the sign.
return append(appendTo, decimalZero)
}
neg := d.Negative != invert
switch d.Form {
case apd.Finite:
// ignore
case apd.Infinite:
if neg {
return append(appendTo, decimalNegativeInfinity)
}
return append(appendTo, decimalInfinity)
case apd.NaN:
if invert {
return append(appendTo, decimalNaNDesc)
}
return append(appendTo, decimalNaN)
default:
panic(errors.Errorf("unknown form: %s", d.Form))
}
e, m := decimalEandM(d, appendTo[len(appendTo):])
return encodeEandM(appendTo, neg, e, m)
}
// decimalEandM computes and returns the exponent E and mantissa M for d.
//
// The mantissa is a base-100 representation of the value. The exponent E
// determines where to put the decimal point.
//
// Each centimal digit of the mantissa is stored in a byte. If the value of the
// centimal digit is X (hence X>=0 and X<=99) then the byte value will be 2*X+1
// for every byte of the mantissa, except for the last byte which will be
// 2*X+0. The mantissa must be the minimum number of bytes necessary to
// represent the value; trailing X==0 digits are omitted. This means that the
// mantissa will never contain a byte with the value 0x00.
//
// If we assume all digits of the mantissa occur to the right of the decimal
// point, then the exponent E is the power of one hundred by which one must
// multiply the mantissa to recover the original value.
func decimalEandM(d *apd.Decimal, tmp []byte) (int, []byte) {
addedZero := false
if cap(tmp) > 0 {
tmp = tmp[:1]
tmp[0] = '0'
addedZero = true
}
tmp = d.Coeff.Append(tmp, 10)
if !addedZero {
tmp = append(tmp, '0')
copy(tmp[1:], tmp[:len(tmp)-1])
tmp[0] = '0'
}
// The exponent will be the combination of the decimal's exponent, and the
// number of digits in the big.Int.
e10 := int(d.Exponent) + len(tmp[1:])
// Strip off trailing zeros in big.Int's string representation.
for tmp[len(tmp)-1] == '0' {
tmp = tmp[:len(tmp)-1]
}
// Convert the power-10 exponent to a power of 100 exponent.
var e100 int
if e10 >= 0 {
e100 = (e10 + 1) / 2
} else {
e100 = e10 / 2
}
// Strip the leading 0 if the conversion to e100 did not add a multiple of
// 10.
if e100*2 == e10 {
tmp = tmp[1:]
}
// Ensure that the number of digits is even.
if len(tmp)%2 != 0 {
tmp = append(tmp, '0')
}
// Convert the base-10 'b' slice to a base-100 'm' slice. We do this
// conversion in place to avoid an allocation.
m := tmp[:len(tmp)/2]
for i := 0; i < len(tmp); i += 2 {
accum := 10*int(tmp[i]-'0') + int(tmp[i+1]-'0')
// The bytes are encoded as 2n+1.
m[i/2] = byte(2*accum + 1)
}
// The last byte is encoded as 2n+0.
m[len(m)-1]--
return e100, m
}
// encodeEandM encodes the exponent and mantissa, appending the encoding to a byte buffer.
//
// The mantissa m can be stored in the spare capacity of appendTo.
func encodeEandM(appendTo []byte, negative bool, e int, m []byte) []byte {
var buf []byte
if n := len(m) + maxVarintSize + 2; n <= cap(appendTo)-len(appendTo) {
buf = appendTo[len(appendTo) : len(appendTo)+n]
} else {
buf = make([]byte, n)
}
switch {
case e < 0:
return append(appendTo, encodeSmallNumber(negative, e, m, buf)...)
case e >= 0 && e <= 10:
return append(appendTo, encodeMediumNumber(negative, e, m, buf)...)
case e >= 11:
return append(appendTo, encodeLargeNumber(negative, e, m, buf)...)
}
panic("unreachable")
}
// encodeVarExpNumber encodes a Uvarint exponent and mantissa into a buffer;
// only used for small (less than 0) and large (greater than 10) exponents.
//
// If required, the mantissa should already be in ones complement.
//
// The encoding must fit in encInto. The mantissa m can overlap with encInto.
//
// Returns the length-adjusted buffer.
func encodeVarExpNumber(tag byte, expAscending bool, exp uint64, m []byte, encInto []byte) []byte {
// Because m can overlap with encInto, we must first copy m to the right place
// before modifying encInto.
var n int
if expAscending {
n = EncLenUvarintAscending(exp)
} else {
n = EncLenUvarintDescending(exp)
}
l := 1 + n + len(m)
if len(encInto) < l+1 {
panic("buffer too short")
}
copy(encInto[1+n:], m)
encInto[0] = tag
if expAscending {
EncodeUvarintAscending(encInto[1:1], exp)
} else {
EncodeUvarintDescending(encInto[1:1], exp)
}
encInto[l] = decimalTerminator
return encInto[:l+1]
}
// encodeSmallNumber encodes the exponent and mantissa into a buffer; only used
// when the exponent is negative. See encodeVarExpNumber.
func encodeSmallNumber(negative bool, e int, m []byte, encInto []byte) []byte {
if negative {
onesComplement(m)
return encodeVarExpNumber(decimalNegSmall, true, uint64(-e), m, encInto)
}
return encodeVarExpNumber(decimalPosSmall, false, uint64(-e), m, encInto)
}
// encodeLargeNumber encodes the exponent and mantissa into a buffer; only used
// when the exponent is larger than 10. See encodeVarExpNumber.
func encodeLargeNumber(negative bool, e int, m []byte, encInto []byte) []byte {
if negative {
onesComplement(m)
return encodeVarExpNumber(decimalNegLarge, false, uint64(e), m, encInto)
}
return encodeVarExpNumber(decimalPosLarge, true, uint64(e), m, encInto)
}
// encodeMediumNumber encodes the exponent and mantissa into a buffer, only used
// when the exponent is in [0, 10].
//
// The encoding must fit in encInto. The mantissa m can overlap with encInto.
//
// Returns the length-adjusted buffer.
func encodeMediumNumber(negative bool, e int, m []byte, encInto []byte) []byte {
l := 1 + len(m)
if len(encInto) < l+1 {
panic("buffer too short")
}
// Because m can overlap with encInto, we must first copy m to the right place
// before modifying encInto.
copy(encInto[1:], m)
if negative {
encInto[0] = decimalNegMedium - byte(e)
onesComplement(encInto[1:l])
} else {
encInto[0] = decimalPosMedium + byte(e)
}
encInto[l] = decimalTerminator
return encInto[:l+1]
}
// DecodeDecimalAscending returns the remaining byte slice after decoding and the decoded
// decimal from buf.
func DecodeDecimalAscending(buf []byte, tmp []byte) ([]byte, apd.Decimal, error) {
return decodeDecimal(buf, tmp, false)
}
// DecodeDecimalDescending decodes decimals encoded with EncodeDecimalDescending.
func DecodeDecimalDescending(buf []byte, tmp []byte) ([]byte, apd.Decimal, error) {
return decodeDecimal(buf, tmp, true)
}
func decodeDecimal(buf []byte, tmp []byte, invert bool) ([]byte, apd.Decimal, error) {
// Handle the simplistic cases first.
switch buf[0] {
case decimalNaN, decimalNaNDesc:
return nil, apd.Decimal{Form: apd.NaN}, nil
case decimalInfinity:
return nil, apd.Decimal{Form: apd.Infinite, Negative: invert}, nil
case decimalNegativeInfinity:
return nil, apd.Decimal{Form: apd.Infinite, Negative: !invert}, nil
case decimalZero:
return buf[1:], apd.Decimal{}, nil
}
tmp = tmp[len(tmp):cap(tmp)]
switch {
case buf[0] == decimalNegLarge:
// Negative large.
e, m, r, tmp2, err := decodeLargeNumber(true, buf, tmp)
if err != nil {
return nil, apd.Decimal{}, err
}
return r, makeDecimalFromMandE(!invert, e, m, tmp2), nil
case buf[0] > decimalNegLarge && buf[0] <= decimalNegMedium:
// Negative medium.
e, m, r, tmp2, err := decodeMediumNumber(true, buf, tmp)
if err != nil {
return nil, apd.Decimal{}, err
}
return r, makeDecimalFromMandE(!invert, e, m, tmp2), nil
case buf[0] == decimalNegSmall:
// Negative small.
e, m, r, tmp2, err := decodeSmallNumber(true, buf, tmp)
if err != nil {
return nil, apd.Decimal{}, err
}
return r, makeDecimalFromMandE(!invert, e, m, tmp2), nil
case buf[0] == decimalPosLarge:
// Positive large.
e, m, r, tmp2, err := decodeLargeNumber(false, buf, tmp)
if err != nil {
return nil, apd.Decimal{}, err
}
return r, makeDecimalFromMandE(invert, e, m, tmp2), nil
case buf[0] >= decimalPosMedium && buf[0] < decimalPosLarge:
// Positive medium.
e, m, r, tmp2, err := decodeMediumNumber(false, buf, tmp)
if err != nil {
return nil, apd.Decimal{}, err
}
return r, makeDecimalFromMandE(invert, e, m, tmp2), nil
case buf[0] == decimalPosSmall:
// Positive small.
e, m, r, tmp2, err := decodeSmallNumber(false, buf, tmp)
if err != nil {
return nil, apd.Decimal{}, err
}
return r, makeDecimalFromMandE(invert, e, m, tmp2), nil
default:
return nil, apd.Decimal{}, errors.Errorf("unknown prefix of the encoded byte slice: %q", buf)
}
}
// getDecimalLen returns the length of an encoded decimal.
func getDecimalLen(buf []byte) (int, error) {
m := buf[0]
p := 1
if m < decimalNaN || m > decimalNaNDesc {
panic(fmt.Errorf("invalid tag %d", m))
}
switch m {
case decimalNaN, decimalNegativeInfinity, decimalNaNDesc, decimalInfinity, decimalZero:
return 1, nil
case decimalNegLarge, decimalNegSmall, decimalPosLarge, decimalPosSmall:
// Skip the varint exponent.
l, err := getVarintLen(buf[p:])
if err != nil {
return 0, err
}
p += l
}
idx, err := findDecimalTerminator(buf[p:])
if err != nil {
return 0, err
}
return p + idx + 1, nil
}
// makeDecimalFromMandE reconstructs the decimal from the mantissa M and
// exponent E.
func makeDecimalFromMandE(negative bool, e int, m []byte, tmp []byte) apd.Decimal {
// ±dddd.
b := tmp[:0]
if n := len(m)*2 + 1; cap(b) < n {
b = make([]byte, 0, n)
}
for i, v := range m {
t := int(v)
if i == len(m) {
t--
}
t /= 2
b = append(b, byte(t/10)+'0', byte(t%10)+'0')
}
if b[len(b)-1] == '0' {
b = b[:len(b)-1]
}
exp := 2*e - len(b)
dec := apd.Decimal{
Exponent: int32(exp),
}
// We unsafely convert the []byte to a string to avoid the usual allocation
// when converting to a string.
s := *(*string)(unsafe.Pointer(&b))
_, ok := dec.Coeff.SetString(s, 10)
if !ok {
panic(fmt.Sprintf("could not set big.Int's string value: %q", s))
}
dec.Negative = negative
return dec
}
// findDecimalTerminator finds the decimalTerminator in the given slice.
func findDecimalTerminator(buf []byte) (int, error) {
if idx := bytes.IndexByte(buf, decimalTerminator); idx != -1 {
return idx, nil
}
return -1, errors.Errorf("did not find terminator %#x in buffer %#x", decimalTerminator, buf)
}
func decodeSmallNumber(
negative bool, buf []byte, tmp []byte,
) (e int, m []byte, rest []byte, newTmp []byte, err error) {
var ex uint64
var r []byte
if negative {
r, ex, err = DecodeUvarintAscending(buf[1:])
} else {
r, ex, err = DecodeUvarintDescending(buf[1:])
}
if err != nil {
return 0, nil, nil, nil, err
}
idx, err := findDecimalTerminator(r)
if err != nil {
return 0, nil, nil, nil, err
}
m = r[:idx]
if negative {
var mCpy []byte
if k := len(m); k <= len(tmp) {
mCpy = tmp[:k]
tmp = tmp[k:]
} else {
mCpy = make([]byte, k)
}
copy(mCpy, m)
onesComplement(mCpy)
m = mCpy
}
return int(-ex), m, r[idx+1:], tmp, nil
}
func decodeMediumNumber(
negative bool, buf []byte, tmp []byte,
) (e int, m []byte, rest []byte, newTmp []byte, err error) {
idx, err := findDecimalTerminator(buf[1:])
if err != nil {
return 0, nil, nil, nil, err
}
m = buf[1 : idx+1]
if negative {
e = int(decimalNegMedium - buf[0])
var mCpy []byte
if k := len(m); k <= len(tmp) {
mCpy = tmp[:k]
tmp = tmp[k:]
} else {
mCpy = make([]byte, k)
}
copy(mCpy, m)
onesComplement(mCpy)
m = mCpy
} else {
e = int(buf[0] - decimalPosMedium)
}
return e, m, buf[idx+2:], tmp, nil
}
func decodeLargeNumber(
negative bool, buf []byte, tmp []byte,
) (e int, m []byte, rest []byte, newTmp []byte, err error) {
var ex uint64
var r []byte
if negative {
r, ex, err = DecodeUvarintDescending(buf[1:])
} else {
r, ex, err = DecodeUvarintAscending(buf[1:])
}
if err != nil {
return 0, nil, nil, nil, err
}
idx, err := findDecimalTerminator(r)
if err != nil {
return 0, nil, nil, nil, err
}
m = r[:idx]
if negative {
var mCpy []byte
if k := len(m); k <= len(tmp) {
mCpy = tmp[:k]
tmp = tmp[k:]
} else {
mCpy = make([]byte, k)
}
copy(mCpy, m)
onesComplement(mCpy)
m = mCpy
}
return int(ex), m, r[idx+1:], tmp, nil
}
// EncodeNonsortingDecimal returns the resulting byte slice with the
// encoded decimal appended to b. The encoding is limited compared to
// standard encodings in this package in that
// - It will not sort lexicographically
// - It does not encode its length or terminate itself, so decoding
// functions must be provided the exact encoded bytes
//
// The encoding assumes that any number can be written as ±0.xyz... * 10^exp,
// where xyz is a digit string, x != 0, and the last decimal in xyz is also
// not 0.
//
// The encoding uses its first byte to split decimals into 7 distinct
// ordered groups (no NaN or Infinity support yet). The groups can
// be seen in encoding.go's const definition. Following this, the
// absolute value of the exponent of the decimal (as defined above)
// is encoded as an unsigned varint. Second, the absolute value of
// the digit string is added as a big-endian byte slice.
//
// All together, the encoding looks like:
// <marker><uvarint exponent><big-endian encoded big.Int>.
//
// The markers are shared with the sorting decimal encoding as follows:
// decimalNaN -> decimalNaN
// decimalNegativeInfinity -> decimalNegativeInfinity
// decimalNegLarge -> decimalNegValPosExp
// decimalNegMedium -> decimalNegValZeroExp
// decimalNegSmall -> decimalNegValNegExp
// decimalZero -> decimalZero
// decimalPosSmall -> decimalPosValNegExp
// decimalPosMedium -> decimalPosValZeroExp
// decimalPosLarge -> decimalPosValPosExp
// decimalInfinity -> decimalInfinity
// decimalNaNDesc -> decimalNaNDesc
//
func EncodeNonsortingDecimal(b []byte, d *apd.Decimal) []byte {
if d.IsZero() && !d.Negative {
// Negative zero will use the decimalNegLarge encoding below.
return append(b, decimalZero)
}
neg := d.Negative
switch d.Form {
case apd.Finite:
// ignore
case apd.Infinite:
if neg {
return append(b, decimalNegativeInfinity)
}
return append(b, decimalInfinity)
case apd.NaN:
return append(b, decimalNaN)
default:
panic(errors.Errorf("unknown form: %s", d.Form))
}
// Determine the exponent of the decimal, with the
// exponent defined as .xyz * 10^exp.
nDigits := int(d.NumDigits())
e := nDigits + int(d.Exponent)
bNat := d.Coeff.Bits()
var buf []byte
if n := UpperBoundNonsortingDecimalSize(d); n <= cap(b)-len(b) {
// We append the marker directly to the input buffer b below, so
// we are off by 1 for each of these, which explains the adjustments.
buf = b[len(b)+1 : len(b)+1]
} else {
buf = make([]byte, 0, n-1)
}
switch {
case neg && e > 0:
b = append(b, decimalNegLarge)
buf = encodeNonsortingDecimalValue(uint64(e), bNat, buf)
return append(b, buf...)
case neg && e == 0:
b = append(b, decimalNegMedium)
buf = encodeNonsortingDecimalValueWithoutExp(bNat, buf)
return append(b, buf...)
case neg && e < 0:
b = append(b, decimalNegSmall)
buf = encodeNonsortingDecimalValue(uint64(-e), bNat, buf)
return append(b, buf...)
case !neg && e < 0:
b = append(b, decimalPosSmall)
buf = encodeNonsortingDecimalValue(uint64(-e), bNat, buf)
return append(b, buf...)
case !neg && e == 0:
b = append(b, decimalPosMedium)
buf = encodeNonsortingDecimalValueWithoutExp(bNat, buf)
return append(b, buf...)
case !neg && e > 0:
b = append(b, decimalPosLarge)
buf = encodeNonsortingDecimalValue(uint64(e), bNat, buf)
return append(b, buf...)
}
panic("unreachable")
}
// encodeNonsortingDecimalValue encodes the absolute value of a decimal's
// exponent and slice of digit bytes into buf, returning the populated buffer
// after encoding. The function first encodes the absolute value of a
// decimal's exponent as an unsigned varint. Following this, it copies the
// decimal's big-endian digits themselves into the buffer.
func encodeNonsortingDecimalValue(exp uint64, digits []big.Word, buf []byte) []byte {
// Encode the exponent using a Uvarint.
buf = EncodeUvarintAscending(buf, exp)
// Encode the digits into the end of the byte slice.
return copyWords(buf, digits)
}
func encodeNonsortingDecimalValueWithoutExp(digits []big.Word, buf []byte) []byte {
// Encode the digits into the end of the byte slice.
return copyWords(buf, digits)
}
// DecodeNonsortingDecimal returns the decoded decimal from buf encoded with
// EncodeNonsortingDecimal. buf is assumed to contain only the encoded decimal,
// as the function does not know from the encoding itself what the length
// of the encoded value is.
func DecodeNonsortingDecimal(buf []byte, tmp []byte) (apd.Decimal, error) {
var dec apd.Decimal
switch buf[0] {
case decimalNaN:
dec.Form = apd.NaN
return dec, nil
case decimalNegativeInfinity:
dec.Form = apd.Infinite
dec.Negative = true
return dec, nil
case decimalInfinity:
dec.Form = apd.Infinite
return dec, nil
case decimalZero:
return dec, nil
}
dec.Form = apd.Finite
switch {
case buf[0] == decimalNegLarge:
if err := decodeNonsortingDecimalValue(&dec, false, buf[1:], tmp); err != nil {
return apd.Decimal{}, err
}
dec.Negative = true
return dec, nil
case buf[0] == decimalNegMedium:
decodeNonsortingDecimalValueWithoutExp(&dec, buf[1:], tmp)
dec.Negative = true
return dec, nil
case buf[0] == decimalNegSmall:
if err := decodeNonsortingDecimalValue(&dec, true, buf[1:], tmp); err != nil {
return apd.Decimal{}, err
}
dec.Negative = true
return dec, nil
case buf[0] == decimalPosSmall:
if err := decodeNonsortingDecimalValue(&dec, true, buf[1:], tmp); err != nil {
return apd.Decimal{}, err
}
return dec, nil
case buf[0] == decimalPosMedium:
decodeNonsortingDecimalValueWithoutExp(&dec, buf[1:], tmp)
return dec, nil
case buf[0] == decimalPosLarge:
if err := decodeNonsortingDecimalValue(&dec, false, buf[1:], tmp); err != nil {
return apd.Decimal{}, err
}
return dec, nil
default:
return apd.Decimal{}, errors.Errorf("unknown prefix of the encoded byte slice: %q", buf)
}
}
func decodeNonsortingDecimalValue(dec *apd.Decimal, negExp bool, buf, tmp []byte) error {
// Decode the exponent.
buf, e, err := DecodeUvarintAscending(buf)
if err != nil {
return err
}
if negExp {
e = -e
}
bi := &dec.Coeff
bi.SetBytes(buf)
// Set the decimal's scale.
nDigits := int(apd.NumDigits(bi))
exp := int(e) - nDigits
dec.Exponent = int32(exp)
return nil
}
func decodeNonsortingDecimalValueWithoutExp(dec *apd.Decimal, buf, tmp []byte) {
bi := &dec.Coeff
bi.SetBytes(buf)
// Set the decimal's scale.
nDigits := apd.NumDigits(bi)
dec.Exponent = -int32(nDigits)
}
// UpperBoundNonsortingDecimalSize returns the upper bound number of bytes
// that the decimal will need for the non-sorting encoding.
func UpperBoundNonsortingDecimalSize(d *apd.Decimal) int {
// Makeup of upper bound size:
// - 1 byte for the prefix
// - maxVarintSize for the exponent
// - WordLen for the big.Int bytes
return 1 + maxVarintSize + WordLen(d.Coeff.Bits())
}
// upperBoundNonsortingDecimalUnscaledSize is the same as
// UpperBoundNonsortingDecimalSize but for a decimal with the given unscaled
// length. The upper bound here may not be as tight as the one returned by
// UpperBoundNonsortingDecimalSize.
func upperBoundNonsortingDecimalUnscaledSize(unscaledLen int) int {
// The number of digits needed to represent a base 10 number of length
// unscaledLen in base 2.
unscaledLenBase2 := float64(unscaledLen) * math.Log2(10)
unscaledLenBase2Words := math.Ceil(unscaledLenBase2 / 8 / float64(bigWordSize))
unscaledLenWordRounded := int(unscaledLenBase2Words) * bigWordSize
// Makeup of upper bound size:
// - 1 byte for the prefix
// - maxVarintSize for the exponent
// - unscaledLenWordRounded for the big.Int bytes
return 1 + maxVarintSize + unscaledLenWordRounded
}
// Taken from math/big/arith.go.
const bigWordSize = int(unsafe.Sizeof(big.Word(0)))
// WordLen returns the size in bytes of the given array of Words.
func WordLen(nat []big.Word) int {
return len(nat) * bigWordSize
}
// copyWords was adapted from math/big/nat.go. It writes the value of
// nat into buf using big-endian encoding. len(buf) must be >= len(nat)*bigWordSize.
func copyWords(buf []byte, nat []big.Word) []byte {
// Omit leading zeros from the resulting byte slice, which is both
// safe and exactly what big.Int.Bytes() does. See big.nat.setBytes()
// and big.nat.norm() for how this is normalized on decoding.
leading := true
for w := len(nat) - 1; w >= 0; w-- {
d := nat[w]
for j := bigWordSize - 1; j >= 0; j-- {
by := byte(d >> (8 * big.Word(j)))
if by == 0 && leading {
continue
}
leading = false
buf = append(buf, by)
}
}
return buf
}
此处可能存在不合适展示的内容,页面不予展示。您可通过相关编辑功能自查并修改。
如您确认内容无涉及 不当用语 / 纯广告导流 / 暴力 / 低俗色情 / 侵权 / 盗版 / 虚假 / 无价值内容或违法国家有关法律法规的内容,可点击提交进行申诉,我们将尽快为您处理。